• Laser & Optoelectronics Progress
  • Vol. 56, Issue 16, 162803 (2019)
Jianlin Wang1, Xiaoqi Lü1、2、*, Ming Zhang1, and Jing Li1
Author Affiliations
  • 1 Inner Mongolia Key Laboratory of Pattern Recognition and Intelligent Image Processing, School of Information Engineering, Inner Mongolia University of Science and Technology, Baotou, Inner Mongolia 0 14010, China
  • 2 School of Information Engineering, Inner Mongolia University of Technology, Hohhot, Inner Mongolia 0 10051, China
  • show less
    DOI: 10.3788/LOP56.162803 Cite this Article Set citation alerts
    Jianlin Wang, Xiaoqi Lü, Ming Zhang, Jing Li. Remote Sensing Image Ship Detection Based on Improved R-FCN[J]. Laser & Optoelectronics Progress, 2019, 56(16): 162803 Copy Citation Text show less
    Diagram of R-FCN network structure
    Fig. 1. Diagram of R-FCN network structure
    Diagram of RoI
    Fig. 2. Diagram of RoI
    Residual structure
    Fig. 3. Residual structure
    Improved residual block structure
    Fig. 4. Improved residual block structure
    Splitting diagrams
    Fig. 5. Splitting diagrams
    Ship detection results of Sentinel-1 images. (a) Picture 1 to be tested; (b) picture 2 to be tested; (c) picture 3 to be tested; (d) original R-FCN detection result of Fig. 6(a); (e) original R-FCN detection result of Fig. 6(b); (f) original R-FCN detection results of Fig. 6(c); (g) detection result of improved Faster R-FCN for Fig. 6(a); (h) detection result of improved Faster R-FCN for Fig. 6(b); (i) detection result of improved Faster R-FCN for Fig. 6(c); (j) detection result of improved R-FC
    Fig. 6. Ship detection results of Sentinel-1 images. (a) Picture 1 to be tested; (b) picture 2 to be tested; (c) picture 3 to be tested; (d) original R-FCN detection result of Fig. 6(a); (e) original R-FCN detection result of Fig. 6(b); (f) original R-FCN detection results of Fig. 6(c); (g) detection result of improved Faster R-FCN for Fig. 6(a); (h) detection result of improved Faster R-FCN for Fig. 6(b); (i) detection result of improved Faster R-FCN for Fig. 6(c); (j) detection result of improved R-FC
    Test results
    Fig. 7. Test results
    Ship detection results of GF-3 image. (a) Picture 1 to be tested; (b) picture 2 to be tested; (c) picture 3 to be tested; (d) detection result of improved R-FCN for Fig. 8(a); (e) detection result of improved R-FCN for Fig. 8(b); (f) detection result of improved R-FCN for Fig. 8(c)
    Fig. 8. Ship detection results of GF-3 image. (a) Picture 1 to be tested; (b) picture 2 to be tested; (c) picture 3 to be tested; (d) detection result of improved R-FCN for Fig. 8(a); (e) detection result of improved R-FCN for Fig. 8(b); (f) detection result of improved R-FCN for Fig. 8(c)
    ParameterValue
    Image size /(pixel×pixel)156×156256×256500×500800×8001000×10001200×12001500×1500
    Ndetection29292929282625
    Nmiss0000134
    Atime /s0.090.090.110.110.130.140.16
    Table 1. Statistic table of ship detection results of images with different sizes
    MethodImageNtargetNdetectionNfalseNmissPPrecise /%RRecall /%
    R-FCNSentinel-11243279978.1220.16
    GF-3822796466.6721.95
    Faster-RCNNSentinel-112410641296.2382.25
    GF-3827641094.7487.80
    Improved R-FCNSentinel-11241183997.4692.74
    GF-382762897.3790.24
    Table 2. Statistical table of ship detection results
    Jianlin Wang, Xiaoqi Lü, Ming Zhang, Jing Li. Remote Sensing Image Ship Detection Based on Improved R-FCN[J]. Laser & Optoelectronics Progress, 2019, 56(16): 162803
    Download Citation